Have you considered shifting the os to win10 for some tests ( with a 2ndary spare boot drive to use as test ), to rule out the fact that win11 could be in part responsible ?
Unfortunately one of the RTXs is now shelved, and the system still fails to reboot due to the no-memory failures I mentioned above. A long press on the power button doesn't seem to do anything, so I'm replugging the power cord in order to start the workstation. Hopefully a BIOS update will solve the no-memory issue.
The system once again became unusable. Significant GUI painting and mouse/keyboard issues. No luck with reboots or swapped RTX cards. No pending OS updates or Dell Command Updates.
Thinking I had no choice but to reinstall Windows 11, both RTX cards were reinstalled per factory settings in preparation for an OS rebuild.
I visited the Product Support Portal to grab an OS image for a reinstallation. (BTW Support Assist via the BIOS wouldn't permit an OS reset from there.) As luck would have it, there was both a critical BIOS (1.0.26) and NVidia (31.0.15.2886, A00) update (on the Support Portal) dated coincidentally today, 2023-06-13. After applying the BIOS update alone, all the GUI issues disappeared including a much improved POST time. Thereafter, I applied the NVidia update. Performance remained solid. Fingers crossed the GUI issues are resolved. I have yet to determine if I'll get clean reboots without the no-memory issue.
The PassMark results are once again similar to the unboxing results. Despite having 128GB DDR4 3200MHz memory, the Memory Mark scores low given the known AMD latency factors. But the CPU & GPU(s) are where they should be. The Disk results are being pulled down by 8TB of secondary spinning disk storage. And finally with the aforementioned BIOS update, the 2D Mark jumped from the teletype spectrum to a reasonable GUI metric.
Getting there, thanks! The frequent no-memory on reboot issue may still exist. This along with arbitrary BSODs occurred since I received the unit, but the BSODs tapered off over time.
A short 24 hour reprieve, but the GUI symptoms returned after the following Windows 11 updates: KB5027231, KB5027119
The 2D Graphics Mark at 8% vs the world. Not good.
Attempted to apply a Restore Point saved just prior to the updates, but the system hangs on reboot which requires replugging the power cord. Tried a second time with similar reboot issues. The OS reported Restore Point failures both times.
At this point, I'm not sure about the cause of the sluggish 2D. Yesterday I thought a BIOS update solved the issue, today it feels like the OS update triggered the symptoms, or maybe the hardware is simply failing unpredictably on reboots. And the only reboots the system completes are hard (replugged) boots. Soft reboots might succeed on rare occasion.
And while plugging the power cord, I heard the following fan noise out of the PSU: PSU Fan Noise (link points to an audio file shared via Dropbox) The system is in a server room, so there's an ambient hum from the other servers. The noise is apparent on a ~10 second cycle.
I'm going by memory about some articles I read in the past month, but I'm pretty sure that they were planning some microcode update/mitigation added to a windows update
The fan noise could be a coincidental fan issue ( with the ball bearing / motor ) requiring a warranty change of the fan (being in the psu, I guess a change of the psu itself), or be induced by the update that somehow affected the power regulation part of the firmware ( ? I'm speculating ).
Personally I still think that it's caused by windows 11
Had a lengthy but productive Pro-Support call yesterday. The tech was quite knowledgeable and given remote access to the PC. The symptoms were sporadic but fortunately occurred during the call. I'm expecting on-site support early next week. I'll provide an update thereafter.
I'm embarrassed to mention how much I spent on this machine. Aside from the frequent no-memory failures (S6, STO) on reboot, I'm now getting a 4% on repeated 2D Passmark tests. The workstation is outfitted with two RTX A4500s. The interactivity on the desktop is impossible. Drivers were regularly kept up to date via Dell Command Update. Windows 11. Every possible Dell diagnostic that I could find reported healthy.
Purchased September 2022, commissioned 3 months later in December.
Glad I didn't decommission my 2017 Precision 7910.
I regret being an early adopter of this AMD configuration. And the hassles I have to look forward to: OS reinstallation and the joy of dealing with Support.
If you're serious about your engineering hours, go with the proven XEON configurations.
The 7865 is a beautiful machine and extremely performant when the stars are aligned, but all that's meaningless when remote reboots fail and the UI feels like VNC on a headless Linux server. I have a $1k Raritan lined up for remote power control -- the reason I asked this question back in March: Dell-Precision-7865-Remote-power-switch-connector
Have you tried running any specific diagnostics or stress tests to pinpoint the root cause of the frequent no-memory failures and low 2D Passmark test results on your workstation with two RTX A4500s, and if so, were there any notable findings that could help identify the underlying issue?
mazzinia_
6 Professor
•
1.5K Posts
0
June 5th, 2023 02:00
Sorry to hear.
Have you considered shifting the os to win10 for some tests ( with a 2ndary spare boot drive to use as test ), to rule out the fact that win11 could be in part responsible ?
williambyrne
1 Rookie
•
11 Posts
0
June 5th, 2023 11:00
That's a good idea, thanks. But I pulled one of the two RTX A4500s which helped.
Similar issue reported here: Bad Graphics in Windows 11
Unfortunately one of the RTXs is now shelved, and the system still fails to reboot due to the no-memory failures I mentioned above. A long press on the power button doesn't seem to do anything, so I'm replugging the power cord in order to start the workstation. Hopefully a BIOS update will solve the no-memory issue.
mazzinia_
6 Professor
•
1.5K Posts
0
June 5th, 2023 11:00
Maybe playing with the bios settings for the gpu slot or such ?
williambyrne
1 Rookie
•
11 Posts
0
June 13th, 2023 20:00
Update 2023-06-13
The system once again became unusable. Significant GUI painting and mouse/keyboard issues. No luck with reboots or swapped RTX cards. No pending OS updates or Dell Command Updates.
Thinking I had no choice but to reinstall Windows 11, both RTX cards were reinstalled per factory settings in preparation for an OS rebuild.
I visited the Product Support Portal to grab an OS image for a reinstallation. (BTW Support Assist via the BIOS wouldn't permit an OS reset from there.) As luck would have it, there was both a critical BIOS (1.0.26) and NVidia (31.0.15.2886, A00) update (on the Support Portal) dated coincidentally today, 2023-06-13. After applying the BIOS update alone, all the GUI issues disappeared including a much improved POST time. Thereafter, I applied the NVidia update. Performance remained solid. Fingers crossed the GUI issues are resolved. I have yet to determine if I'll get clean reboots without the no-memory issue.
The PassMark results are once again similar to the unboxing results. Despite having 128GB DDR4 3200MHz memory, the Memory Mark scores low given the known AMD latency factors. But the CPU & GPU(s) are where they should be. The Disk results are being pulled down by 8TB of secondary spinning disk storage. And finally with the aforementioned BIOS update, the 2D Mark jumped from the teletype spectrum to a reasonable GUI metric.
mazzinia_
6 Professor
•
1.5K Posts
0
June 14th, 2023 04:00
Happy to hear that it's getting solved
williambyrne
1 Rookie
•
11 Posts
0
June 14th, 2023 09:00
Getting there, thanks! The frequent no-memory on reboot issue may still exist. This along with arbitrary BSODs occurred since I received the unit, but the BSODs tapered off over time.
williambyrne
1 Rookie
•
11 Posts
0
June 14th, 2023 20:00
Update 2023-06-14
A short 24 hour reprieve, but the GUI symptoms returned after the following Windows 11 updates: KB5027231, KB5027119
The 2D Graphics Mark at 8% vs the world. Not good.
Attempted to apply a Restore Point saved just prior to the updates, but the system hangs on reboot which requires replugging the power cord. Tried a second time with similar reboot issues. The OS reported Restore Point failures both times.
At this point, I'm not sure about the cause of the sluggish 2D. Yesterday I thought a BIOS update solved the issue, today it feels like the OS update triggered the symptoms, or maybe the hardware is simply failing unpredictably on reboots. And the only reboots the system completes are hard (replugged) boots. Soft reboots might succeed on rare occasion.
And while plugging the power cord, I heard the following fan noise out of the PSU: PSU Fan Noise (link points to an audio file shared via Dropbox) The system is in a server room, so there's an ambient hum from the other servers. The noise is apparent on a ~10 second cycle.
mazzinia_
6 Professor
•
1.5K Posts
0
June 15th, 2023 01:00
...
I'm going by memory about some articles I read in the past month, but I'm pretty sure that they were planning some microcode update/mitigation added to a windows update
The fan noise could be a coincidental fan issue ( with the ball bearing / motor ) requiring a warranty change of the fan (being in the psu, I guess a change of the psu itself), or be induced by the update that somehow affected the power regulation part of the firmware ( ? I'm speculating ).
Personally I still think that it's caused by windows 11
williambyrne
1 Rookie
•
11 Posts
0
June 16th, 2023 17:00
Update 2023-06-16
A reassuring Pro-Support case currently underway.
For reference, the No Memory POST code (link points to video file shared via Dropbox)
williambyrne
1 Rookie
•
11 Posts
0
June 16th, 2023 17:00
Agreed regarding a coincidental PSU issue.
Had a lengthy but productive Pro-Support call yesterday. The tech was quite knowledgeable and given remote access to the PC. The symptoms were sporadic but fortunately occurred during the call. I'm expecting on-site support early next week. I'll provide an update thereafter.
I very much appreciate your input!
mazzinia_
6 Professor
•
1.5K Posts
0
June 17th, 2023 03:00
Good luck, it's definitely an annoying issue (while at least the fan is just "cosmetic", at that point... had it happen to my unit, too)
MarvinOcean
2 Posts
0
August 5th, 2023 04:00
Have you tried running any specific diagnostics or stress tests to pinpoint the root cause of the frequent no-memory failures and low 2D Passmark test results on your workstation with two RTX A4500s, and if so, were there any notable findings that could help identify the underlying issue?